RE: [U-Boot-Users] what is wrong when calloc return bogus?
You say loading.....are you starting from flash, then relocating to the middle or RAM _or_ are you using something to download a binary image into the middle of memory, then trying to execute it?
There aren't that many files to configure, it should be straight forward to find. JTAG is a wonderful thing, you should take the time to get it working...assuming your hardware has wired it up correctly.
Regards,
Richard W.
-----Original Message----- From: Matthew S. McClintock [mailto:mattsm@arlut.utexas.edu] Sent: Thursday, November 13, 2003 11:45 AM To: Woodruff, Richard Cc: Holger Schurig; u-boot-users@lists.sourceforge.net Subject: RE: [U-Boot-Users] what is wrong when calloc return bogus?
Ahh ok, we are using that code you mention. I thought for a second there was another CVS branch or something for the ARM925 tree which did not make sense but I thought I would ask. Our board is still failing at malloc. Our board is _very_ similar to the Innovator 1510 and we are using most of that code as a starting point. We are loading u-boot right in the middle of our memory so it should have plenty of space above and below the u-boot code for malloc. I don't have a JTAG debugger working well with our board so I have tracked down the problems I have mentioned using printf statements. The area of memory set aside for malloc is valid unused memory space I just can not quite see why malloc is failing. If you have any ideas, feel free to let me know otherwise I will keep debugging it.
Thanks, Matthew
This is our output without our printf statements, however its not very helpful...
U-Boot 1.0.1 (Nov 11 2003 - 11:29:53)
U-Boot code: 11000000 -> 11014330 BSS: -> 11017FBC RAM Configuration: Bank #0: 10000000 32 MB Flash: 0 kB *** Warning - bad CRC, using default environment
On Thu, 2003-11-13 at 11:29, Woodruff, Richard wrote:
U-boot 1.0.0 and before for the last ... ?Six months has
had an OMAP
tree. Its really trivial to port from that to a new target.
There is
a omap1510inn target which is the Innovator. I have a couple tress which I've not yet merged in for custom boards which have a
lot more
functionality, but we will do that at some point, probably
closer to
the time they actually go into production.
Cpu/arm925 is for the arm core on the OMAP15xx. I've
sprinkled enough
code in to idle the dsp but not much beyond that. The is a
926 tree
also, that is for the omap 16xx. The CPU core is common in
a lot of
different SOC's, that's what the CPU was broke out separately. It could have been done differently, but the hope is to allow
sharing of
common code bits, not duplicate them.
It seems you must be using an older tree as this kind of
thing is sort
of hard to miss if you have gone though the code.
Regards,
Richard W.
-----Original Message----- From: Matthew S. McClintock [mailto:mattsm@arlut.utexas.edu] Sent: Thursday, November 13, 2003 11:20 AM To: Woodruff, Richard Cc: Holger Schurig; u-boot-users@lists.sourceforge.net Subject: RE: [U-Boot-Users] what is wrong when calloc
return bogus?
On Thu, 2003-11-06 at 17:05, Woodruff, Richard wrote:
-- If you cloned the ARM925 tree, then you should NOT see this failure. Its probably some other kind of misconfiguration.
You mention an ARM925 tree, is there ARM925 tree for u-boot somewhere I am missing? -- Matthew S. McClintock mattsm@arlut.utexas.edu
-- Matthew S. McClintock mattsm@arlut.utexas.edu
Its loading from flash, copying itself to the middle of memory. We are currently using a minimal version of RedBoot that came with the development board to load u-boot to flash. The current setup we have as I understand is JTAG via CodeComposer in windows. I don't have any of this setup on my computer, but my college does so I do not have easy access. Also I have tried to use the JTAG to look at memory with CodeComposer but after u-boot starts on the board the debugging via JTAG no longer seems to respond. I still have not had much time to look over all the code so meanwhile I will keep looking until I figure out what is going on.
Thanks, Matthew
On Thu, 2003-11-13 at 11:53, Woodruff, Richard wrote:
You say loading.....are you starting from flash, then relocating to the middle or RAM _or_ are you using something to download a binary image into the middle of memory, then trying to execute it?
There aren't that many files to configure, it should be straight forward to find. JTAG is a wonderful thing, you should take the time to get it working...assuming your hardware has wired it up correctly.
Regards,
Richard W.
-----Original Message----- From: Matthew S. McClintock [mailto:mattsm@arlut.utexas.edu] Sent: Thursday, November 13, 2003 11:45 AM To: Woodruff, Richard Cc: Holger Schurig; u-boot-users@lists.sourceforge.net Subject: RE: [U-Boot-Users] what is wrong when calloc return bogus?
Ahh ok, we are using that code you mention. I thought for a second there was another CVS branch or something for the ARM925 tree which did not make sense but I thought I would ask. Our board is still failing at malloc. Our board is _very_ similar to the Innovator 1510 and we are using most of that code as a starting point. We are loading u-boot right in the middle of our memory so it should have plenty of space above and below the u-boot code for malloc. I don't have a JTAG debugger working well with our board so I have tracked down the problems I have mentioned using printf statements. The area of memory set aside for malloc is valid unused memory space I just can not quite see why malloc is failing. If you have any ideas, feel free to let me know otherwise I will keep debugging it.
Thanks, Matthew
This is our output without our printf statements, however its not very helpful...
U-Boot 1.0.1 (Nov 11 2003 - 11:29:53)
U-Boot code: 11000000 -> 11014330 BSS: -> 11017FBC RAM Configuration: Bank #0: 10000000 32 MB Flash: 0 kB *** Warning - bad CRC, using default environment
On Thu, 2003-11-13 at 11:29, Woodruff, Richard wrote:
U-boot 1.0.0 and before for the last ... ?Six months has
had an OMAP
tree. Its really trivial to port from that to a new target.
There is
a omap1510inn target which is the Innovator. I have a couple tress which I've not yet merged in for custom boards which have a
lot more
functionality, but we will do that at some point, probably
closer to
the time they actually go into production.
Cpu/arm925 is for the arm core on the OMAP15xx. I've
sprinkled enough
code in to idle the dsp but not much beyond that. The is a
926 tree
also, that is for the omap 16xx. The CPU core is common in
a lot of
different SOC's, that's what the CPU was broke out separately. It could have been done differently, but the hope is to allow
sharing of
common code bits, not duplicate them.
It seems you must be using an older tree as this kind of
thing is sort
of hard to miss if you have gone though the code.
Regards,
Richard W.
-----Original Message----- From: Matthew S. McClintock [mailto:mattsm@arlut.utexas.edu] Sent: Thursday, November 13, 2003 11:20 AM To: Woodruff, Richard Cc: Holger Schurig; u-boot-users@lists.sourceforge.net Subject: RE: [U-Boot-Users] what is wrong when calloc
return bogus?
On Thu, 2003-11-06 at 17:05, Woodruff, Richard wrote:
-- If you cloned the ARM925 tree, then you should NOT see this failure. Its probably some other kind of misconfiguration.
You mention an ARM925 tree, is there ARM925 tree for u-boot somewhere I am missing? -- Matthew S. McClintock mattsm@arlut.utexas.edu
-- Matthew S. McClintock mattsm@arlut.utexas.edu
participants (2)
-
Matthew S. McClintock -
Woodruff, Richard